Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [1] Presented By SARAT KUMAR BEHERA Roll # CS NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY DISTRIBUTED VIDEO STREAMING OVER INTERNET Under the guidance of Mr. Dutikrushna Panda

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [2]  It is defined as streaming video from multiple senders simultaneously to a single receiver.  It is also called a Multiple Sender Video Streaming (MSVS) system.  Example: two or more senders senders sending simultaneously at different rate DISTRIBUTED VIDEO STREAMING S1 S2 S3 Receiver 400kbps 300kbps 100kbps

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [3] ADVANTAGES OF MSVS SYSTEM  MSVS uses aggregated bandwidth to ease bandwidth requirements of video streaming.  MSVS combat loss using path diversity, offering higher stability and higher quality of service.

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [4] ISSUES ON MSVS SYSTEM  Encoding and Decoding  Multilayered Coding  Fine Granularity Scalability(FGS)  Transmission Protocol  Sender selection

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [5] ENCODING AND DECODING Compression is employed to achieve transmission efficiency. Some important streaming requirements are:  Bandwidth  Delay  Loss  VCR  Codec efficiency  Decodable and composable  Bandwidth utilization

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [6] COMPRESSION SCHEME A scalable video encoder compresses a raw video sequence to multiple sub-streams. Two sub-streams are generated, namely the “Base Layer (BL)” and the Enhancement layer (EL). DCT QVLC IQ VLCQVLDIDCTIQ IDCT VLDIQ Base Layer compressed Bit-stream Enhanceme nt Layer Compressed Bit-stream Base layer decoded video Enhanced layer decoded video Raw video

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [7] Fine Granularity Scalability  FGS uses bit-plane coding instead of traditional run-level coding of DCT coefficients to achieve the continuous quality curve.  Bit-plane coding considers each quantized DCT coefficient as a binary number of several bits rather than a decimal integer of a certain value.  For each DCT block we have 64 DCT coefficients zigzagged into an array e.g. 10 DEC 1010 BIN  The number of bit-plane is determined by the largest value of all DCT coefficients.

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [8] TRANSMISSION PROTOCOL  The goal of designing a transmission protocol is to ease constraints like bandwidth, loss rate and latency.  There are two aspects in transmission protocol: –Who to send –What to send

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [9] WHO TO SEND This protocol measures the suitability for a host to be a streaming server is required in this protocol. This measurement may focus on bandwidth, loss, latency or path diversity as:  Maximize bandwidth  Minimize overall loss rate  Minimize latency(delay)  Reduce correlation between paths

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [10] WHAT TO SEND This is divided into three aspects :  How to divide the media data  How to distribute  How to use retransmission  Retransmission from original sender  Retransmission from other senders  Retransmission from dedicated retransmitter

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY The transport protocol is a receiver-driven one in which, the receiver coordinates transmissions from multiple senders based on the information received from the senders. Factors affecting transport protocol  Bandwidth Estimation  Rate Allocation Algorithm  Packet Partition Algorithm (PPA) TRANSPORT PROTOCOL FOR RECEIVER SECTION

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [12] BANDWIDTH ESTIMATION The receiver estimates available bandwidth for each sender based on the TCP-friendly rate control algorithm (TFRC). It calculates the available bandwidth according to the following equation where B=current available TCP-friendly bandwidth between each sender and the receiver T rto =TCP time out R = estimated round-trip time p = estimated loss rate s =TCP segment size in bytes

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [13] RATE ALLOCATION ALGORITHM In this protocol, the receiver computes the optimal sending rate for each sender based on its loss rate and estimated available bandwidth. Let N=total no. of senders L(i,t) =estimated loss rate S(i,t) = estimated sending rate Sreq(t)= the required bit rate for the encoded video

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [14] PACKET PARTITION ALGORITHM  After receiving the control packet from the receiver, each sender immediately decides the next packet in the video stream to be sent, using the packet partition algorithm.  All the senders simultaneously run this algorithm to ensure that no sender sends the same video packet, and also to minimize the probability of packets arriving late at the receiver due to network jitter.  The format of the control packet is shown here(for simplicity, the IP header and sequence number is not shown).

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [15]  S1-S5 are two-byte fields to specify the sending rate in packets/second for each sender.  D1-D5 are one-byte fields to denote the estimated delay from each sender to the receiver.  The Sync field is the starting sequence number that all senders use in the packet partition algorithm to determine the next packet to send, immediately upon receiving the control packet. D1D2D3D4D5S1S2S3S4S5sync

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [16] CONCLUSION A framework for a MSVS is proposed in order to achieve higher throughput, and to increase tolerance, to loss and delay due to congestion. This results in lower distortion than distributed video streaming in a simplistic, non-adaptive fashion. It is speculated that the MSE will be even smaller if the protocol can be extended to allocate video bits based on their importance.

Technical Seminar Presentation Presented by : SARAT KUMAR BEHERA NATIONAL INSTITUTE OF SCIENCE AND TECHNOLOGY [17] Thank You…